MemoMorph

MemoMorph helps university students turn fragmented lecture and group recordings into clear, structured notes by merging overlaps, filling gaps, and showing uncertainty instead of hiding it.

Comment

About MemoMorph

Inspiration

MemoMorph came from a common experience shared by many students: recording lectures and group discussions in bits and pieces.

In reality, students seldom have one complete recording. They often gather voice messages from classmates, partial recordings from their phones, and clips captured at different times. When exams approach, the issue isn’t the lack of information; it’s that everything is scattered, overlapping, and inconsistent.

Most existing tools concentrate on transcription or summarization, but they assume the input is clean. MemoMorph starts with the understanding that academic audio is often messy, and tools should be built accordingly.

What I Learned

Creating MemoMorph showed me that working with AI involves more than just generating answers. It’s about designing the right framework to handle uncertainty.

I learned:

  • How to create AI-assisted workflows that reveal uncertainty instead of hiding it
  • How to translate messy human behavior into clear interaction models
  • The importance of pacing, visual clarity, and trust when designing tools for students
  • That a good demo focuses on guiding understanding rather than showcasing features

Most importantly, I realized that clarity is a design choice and not an automatic outcome of using AI.

How I Built the Project

MemoMorph is designed as a web-based application aimed at rapid iteration and straightforward user experience:

  • Frontend: Vite + React
  • ASR: Whisper, running through a lightweight Express backend
  • AI Reconstruction: Gemini API, accessed directly from the frontend
  • Core Logic:
    • Fragment-level processing
    • Detection of overlaps, duplicates, and complementary content
    • Context-aware reconstruction for lectures and group discussions

Instead of offering a single “final answer,” MemoMorph maintains a transparent process:

  • Original transcript
  • Reconstructed structure
  • Clear markers for inferred or conflicting content

This approach helps students understand how results are generated, not just consume them.

Challenges Faced

One significant challenge was avoiding over-automation.

It was tempting to mask uncertainty and present something that seemed perfectly polished. However, this would break trust, especially in academic settings.

Other challenges included:

  • Designing merge logic that feels intuitive rather than “AI magic”
  • Keeping the interface simple while managing complex underlying logic
  • Crafting a meaningful story within a strict 3-minute demo format

Balancing power, clarity, and honesty proved to be the most difficult—and valuable—aspect of the project.

Closing Thoughts

MemoMorph is not merely a transcription tool. It aims to rethink how students engage with imperfect information.

By treating fragmentation as a significant problem rather than an anomaly, MemoMorph transforms messy academic audio into something students can effectively study.

Share this project:

Updates